Analyse appartements - Shapash

Interpretation des predictions appartements

Project_Information

Author : VotreNom

Description : Rapport Shapash pour appartements

Project_Name : Analyse lightgbm_appart


Model analysis

Model used : LGBMRegressor

Library : lightgbm.sklearn

Library version : 4.6.0

Model parameters :

Parameter key Parameter value
boosting_type gbdt
objective None
num_leaves 95
max_depth 10
learning_rate 0.06774931487278521
n_estimators 294
subsample_for_bin 200000
min_split_gain 0.0
min_child_weight 0.001
min_child_samples 10
subsample 1.0
subsample_freq 0
colsample_bytree 1.0
reg_alpha 0.0
reg_lambda 0.0
random_state None
Parameter key Parameter value
n_jobs None
importance_type split
_Booster
_evals_result {}
_best_score defaultdict(, {})
_best_iteration 0
_other_params {}
_objective regression
class_weight None
_class_weight None
_class_map None
_n_features 56
_n_features_in 56
_classes None
_n_classes -1
fitted_ True

Dataset analysis

Global analysis

Training dataset Prediction dataset
number of features NaN 56
number of observations NaN 2,759
missing values NaN 0
% missing values NaN 0

Univariate analysis

etage - Numeric

Prediction dataset
count 2,759
mean -0.0135
std 0.976
min -0.514
25% -0.514
50% -0.155
75% 0.205
max 17.5

Target analysis

prix_m2_vente - Numeric

Prediction dataset
count 2,759
mean 2,550
std 1,070
min 216
25% 1,710
50% 2,440
75% 3,300
max 7,440

Multivariate analysis


Model explainability

Note : the explainability graphs were generated using the test set only.

Global feature importance plot

Features contribution plots

etage -


Model performance

Univariate analysis of target variable

prix_m2_vente - Numeric

True values Prediction values
count 2,759 2,759
mean 2,550 2,560
std 1,070 930
min 216 508
25% 1,710 1,820
50% 2,440 2,440
75% 3,300 3,230
max 7,440 6,880

Metrics

MAE : 355

R2 : 0.783

MSE : 250,000

MAPE : 0.171

MdAE : 251

Explained Variance : 0.783